In SPECT imaging, data acquisition is usually completed in a single scan. A scan usually obtains between 30 and 128 projection frames are acquired in a 180° to 360° angular range, depending upon the clinical protocols. SPECT often has a low count rate because of low system sensitivity and patient dose compared to other imaging modalities, such as CT. The SPECT scan hence takes a relatively long time, e.g., from minutes to tens of minutes Theoretically, SPECT imaging requires the patient to stay motionless during the entire scan. Any movement by the patient causes data inconsistency which may introduce motion artifacts into the final images.
However, it is very difficult to keep patients motionless during scans, and especially during longer scans. When the patient motion becomes too large, the data may not be medically usable. Patients are generally rescanned when this happens.
For example, for cardiac SPECT, ASNC guidelines specifically requires that if patient motion is larger than a certain amount at any of the projection frames, the patients need to be rescanned.
Wang et al (Wang et al 2005) used multiple acquisition of moderate length to reduce effects of radioactive decay for phantom experiments. Chen et al (Chen et al 2004) used multiple sequential scans (fast fanning) with 1 second per projection for dynamic SPECT using Teboroxime. However, neither of these works could be directly used to handle motion correction in SPECT imaging.
A number of motion correction techniques have been proposed for SPECT imaging. These efforts can be divided into two main categories. One is a hardware approach that tracks patient motion during patient scans using a tracking device and uses the motion information during image reconstruction to correct for the motion. The other analyzes the acquired data after the scan using software, where the data are acquired using the current single but relatively long scan approach.
The hardware approach is ideal for motion tracking and correction for patient studies, but the device can be complicated and expensive.
The software approaches are limited in their ability to correct for sub-frame patient motion. For example, if the scanning time for one frame is 20 seconds, and the patient moves only for the second half (10 seconds) of the acquisition of the frame, unless the data are acquired in list mode, none of the current motion correction techniques will operate properly.